The effect 0f male circumcision on female HIV infection – the evidence
If it becomes obvious mid-way through a randomised trial that a new treatment doesn’t work sufficiently well to make it worth using, stopping early is good practice; we may never get a clear answer, but who cares? But if you’re planning on using the new treatment anyway, you really must finish the trial, and sort out once and for all whether it does harm.
Consider a treatment which affects husbands and wives. Treating one automatically treats the other. Evidence accumulates that it is good for the husband but results are unclear for the wife. It would be bonkers to stop when interim data showed a non-significant harmful trend for wives, on the grounds that the treatment was never going to be good for them, and then go ahead and implement it! But that’s what the authors of the only trial ever to test the effect of male circumcision on HIV transmission to the woman did (click here).
They tested the hypothesis that circumcision of HIV positive men would reduce new cases of HIV in female partners, as part of a larger trial including HIV negative men. When the HIV negative cohort closed, the investigators fretted that “continuation […] in HIV-infected men could result in stigmatization” and “determined that the conditional power to detect 60% efficacy, […] was only 4.9% and recommended that enrolment be closed.”
At that point 17/93 (18%) intervention and 8/70 (11%) control women had become infected. The difference might have occurred by chance (hazard ratio 1.58, 95% CI: 0.68–3.66, p=0.287), but if it was real, wives would surely want to know. The harmful effect is biologically plausible; the foreskin functions as a sleeve within which the penile shaft moves during intercourse to reduce vaginal and penile abrasions. But the trial stopped, and the US, WHO, and the many governments who get funds from those sources, went on to encourage male circumcision, without mentioning the possible increase in female infection risk. For the WHO fact sheet (click here).
Some circumcision advocates (e.g. here) admit that the evidence on male to female transmission is unclear, but conclude “that women will benefit from […] voluntary medical circumcision programmes in the long-term” on the basis of modelling studies like this one (click here).
But the modelling studies ignore even the possibility of increased male to female transmission. For the base case this one assumed 60% effectiveness for men, 80% coverage by 2015, and no post-circumcision behaviour change. They tested the effect of varying all these in a sensitivity analyses but not the effect of any increased male to female transmission. Am I going mad?
Let’s summarise. Randomised trials show that male circumcision reduces female to male HIV transmission by about 60%. The only trial which measured it showed a biologically plausible increase in male to female transmission of 58%, but was stopped before it gave a clear answer. Experts modelled the effect of their programmes assuming the male benefit was real, but ignoring even the possibility of increased male to female infection. On the basis of such models millions of men are circumcised, and millions of uninformed women are put at risk.
It’s enough to turn even me feminist!
By Olena Kalytiak Davis
In the 13th century Francesca da Rimini, stuck in an arranged marriage, shared the story of the knights Lancelot and Galehot’s love triangle with Queen Guinevere, with her husband’s brother Paolo. It struck a cord, and overcome by passion they began a ten-year affair. When her husband finally caught them at it, he murdered them both. In the Inferno Dante meets Francesca and Paolo in the second circle of hell, reserved for the lustful.
Davis calls this poem, one of a series about Francesca, a “shattered sonnet”, and includes it in her latest collection The Poem She Didn’t Write and Other Poems (click here).
that maiden thump was book on floor, but
does it really matter who kissed who
first or then who decided to go further?
lower? faster? naturally, we took
turns on top. now here, now there, and up
and down…once it started no one even thought to think to stop.
so, we have holes inside our souls,
but mustn’t we begin by filling others’?
god gave us lips and hands and parts
that cannot possibly be saved for prayer. nor by.
i will not name name, claim fame by how well
or who I fucked or why, it happens all the time.
and it’s you, white pilgrims, whom next galehot seeks.
fuck. we didn’t read again for weeks.
Olena Kalytiak Davis
Often a vivid story or picture carries more weight than any amount of science or ethics. This blog received more hits for some pictures of circumcision complications (click here) than for any other post. Comments about flawed newborn tongue-tie cutting trials were ignored until I drew parallels with circumcision (click here).
This is why I’m publicising this video clip of a brit milah. There’s no blood, and we don’t see the surgery. Just a few off colour jokes, men chanting over the baby’s cries, and the women’s mixed feelings. It’s been posted by the baby. He wants you to watch. Click here.
Three rules – marry, don’t pay blackmail, and be kind to dogs
Jeremy Thorpe, leader of the tiny UK Liberal party from 1967-76, followed the first rule, twice. His first marriage of convenience in 1968 quashed rumours of rent boys, and his wife’s death in a car crash, ten days after the 1970 general election, sealed the sympathy vote. His second wife, previously married to the Queen’s cousin with a stately home near Ripon, helped the 1973 Liberal bye election victory there. In the 1974 general election the Liberals gained 14 MPs in a hung parliament, and the Prime Minister offered Thorpe a cabinet post. He turned it down. By then he had neglected rule two.
A blackmailing former lover, Norman Scott, had kept making demands, so Thorpe mused that it would be good if he was bumped off. We will never know if he was serious, but a gunman did meet up with Scott on Bodmin moor, and shots were fired. If the bullets had killed Scott, that might have been the end of the matter; the plan was to drop the friendless troublemaker down an old mineshaft. But after the assassin killed Scott’s dog Rinka, the gun had jammed and Scott got away.
Oh dear! The British may forgive a sex scandal, but killing an innocent dog? At his trial for conspiracy to murder, a witness claimed that Thorpe had said: “We’ve got to get rid of [Scott]. It is no worse than shooting a sick dog”. To everyone’ surprise he was acquitted, but at the subsequent election a comedian stood against him for The Dog Lovers Party, and he lost to his Conservative opponent. He lived another 35 years, but he’d broken the third rule and there was no way back.
The “young apologist” brought in to replace the true believing vicar in Cider with Rosie (p222) was my grandfather, Cyril Hodson. My mother Helen was too young to remember much of Laurie Lee, but his redoubtable mother, Mrs Lee, made an impression. So when a route was marked out in the Slad valley to celebrate the author’s centenary (click here) her three sons walked it.
Ignore the marker at Bulls Cross car park, walk 50 yards north on the B470, and turn right down a wooded track to the dam in the valley. A hard pull up the other side takes you into the isolated Dillay Brook valley. Some signs have been vandalised – someone dislikes tourists – but who cares.
It’s a perfect walk.
But the poems set into each marker post aren’t a patch on his two great “coming-of-age” books CWR and As I Walked out One Midsummer Morning. Here’s one.
Behold the apples’ rounded worlds:
juice-green of July rain,
the black polestar of flowers, the rind
mapped with its crimson stain.
The russet, crab and cottage red
burn to the sun’s hot brass,
then drop like sweat from every branch
and bubble in the grass.
They lie as wanton as they fall,
and where they fall and break,
the stallion clamps his crunching jaws,
the starling stabs his beak.
In each plump gourd the cidery bite
of boys’ teeth tears the skin;
the waltzing wasp consumes his share,
the bent worm enters in.
I, with as easy hunger, take
entire my season’s dole;
welcome the ripe, the sweet, the sour,
the hollow and the whole.
So much is wrong; the forced rhyme of “dole”, “the cidery bite/of boys’ teeth” – who bites into a fermenting apple? Was repetition of “fall” unavoidable? Does he really eat hollowed out and sour apples?
In 1936 the young apologist climbed the steeple at nearby Uplands church to place the weather vane on top (click here), and 78 years later we toasted him in The Woolpack’s excellent cider.
And labour induction at term
Labour induction generates strong feelings. Parents who lose a normal baby near term (37-42 weeks) understandably wish someone had induced birth earlier. Natural childbirth enthusiasts counter that induction increases pain, Caesarean sections and medicalises childbirth. Both have a point, although the supposed effect on Caesareans may be overstated (click here).
Today I want to consider another aspect of the debate, the way in which the risk of remaining undelivered varies with gestational age.
Many critics of induction use graphs of perinatal mortality by week (e.g. solid dots above) where the risk only rises after 42 weeks, to imply that induction at say 40 or 41 weeks is illogical because the risk is not raised at these gestations.
But such graphs mislead. They express mortality as deaths per 1000 total births. i.e. the denominator for later deaths includes babies who are already safely delivered. This makes no sense if we are concerned about stillbirth because a delivered baby cannot be stillborn. The correct denominator should be babies still undelivered at the particular gestational age. As this proportion falls after 37 weeks so the risk of stillbirth per undelivered baby starts to rise.
But plotting only stillbirth risk is also misleading because babies may also die after birth from complications of labour or prematurity; delivery at 24 weeks would prevent all stillbirths but would hardly be sensible. The data we need are the rates of all deaths per 1000 undelivered babies by gestational age at delivery – the perinatal risk index (open dots above).
The idea takes a bit of getting your head around. The person who explained it to me was Gordon Smith from Cambridge. He has written about it many times. Click here for his paper, from which I nicked the graph. Or here if you have access problems Gordon Smith perinatal risk index paper.
The fact that 38 weeks is the safest gestation to be born doesn’t automatically mean we should induce everyone then; wrong dates might cause prematurity related problems. But 42 weeks is too late.
For accurately dated pregnancies induction at 39 or 40 weeks would prevent many deaths. If it didn’t increase Caesareans, what’s not to like?
Disappearing OA journals
In the early 1990s a group of obstetricians in six UK hospitals (The UK Amniotomy Group) conducted a randomised trial comparing routine immediate amniotomy, with selective and delayed amniotomy for women in first labour. Policies about amniotomy were hotly debated at the time. Partly funded by a grant from the European Union, our trial remains the world’s largest.
In 1994 it fell to me, as corresponding author, to choose where to submit it. Young and keen, and wanting to be up to date, I sent it to a new online open access journal The Online Journal of Current Clinical Trials and it was accepted and published. Click here for the PubMed abstract, but don’t try finding the full article; the journal folded in 1996 and our paper has vanished without trace.
Fortunately, I had also been able to persuade the British Journal of Obstetrics and Gynaecology (BJOG), a conventional subscription journal, to publish a shortened version in parallel (click here), now the only one available. I’ve learnt my lesson. Beware online only open access journals; they may melt away.
And the effect of routine amniotomy? It shortens labour by about an hour on average, but doesn’t alter any substantive adverse outcomes one way or the other. Obstetricians and midwives have sensibly stopped arguing about it, and follow parents’ wishes.
Remember “free” is not always best. Sometimes it’s better to pay.